8 BULLETIN 832, U. 8. DEPARTMENT OF AGRICULTURE. 



It was found that abrupt changes in the size of a channel that are 

 repeated at short intervals are often alone responsible for the high 

 values of n obtained. 



Figures 1, 3, 5, 7, 9, and 11 show the average cross sections for all 

 of the channels. From these a general idea can be obtained as to 

 the size and shape of the various channels. 



SIDE SLOPES AND BOTTOM. 



Irregularities in the perimeter of the channel are discussed under 

 the two headings, side slopes and bottom, for the reason that the 

 bottom alone is often the chief retarding factor during low stages. 

 Irregularities in the perimeters of channels have a very decided effect 

 upon the value of n. For instance, irregularities in the side slopes 

 of the South Forked Deer River near Jackson, Tenn., were largely 

 responsible for the higher values of n obtained for this course than 

 were obtained for the same channel near Roberts, Tenn., where the 

 side slopes were much more regular (Table 3). The irregularities 

 in the channel at Jackson were due to rough finishing work with the 

 dredge at the time of construction. To smoothness of perimeter 

 may be attributed in part the low values of n obtained for the 

 Allen and Willow Creek channels in western Iowa (Table 4). 



Under this heading descriptions of the different types of soils found 

 in the various channels are given. However, no noticeable difference 

 in the retarding effect of the different soils was detected. All of the 

 soils are of an alluvial nature, the bottoms through which the ditches 

 were dug having been built up by deposits of silt brought down from 

 the hills daring heavy rains. 



CONDITION. 



An effort has been made to describe accurately the condition of the 

 channels at the time the measurements were made. Mention is made 

 of ail sorts of growth, obstructions, and debris — factors which are in 

 many instances responsible for the high values of n obtained. The 

 pronounced effect of vegetation, such as grass, in a channel is clearly 

 shown by the experiments made on three of the channels in North 

 Carolina (Table 5), of the effect of small growth, such as weeds, 

 bushes, and saplings by the experiments on Old Town Creek (Table 

 1), and of the effect of large obstructions, such as logs, fallen trees, 

 and debris by the experiments on the South Forked Deer River at 

 Campbell's levee, near Jackson, Tenn. (Table 3). Whether or not 

 much erosion lias taken place in the channel is also mentioned under 

 this heading. 



